ckbiosintesis

7/29/2019 CKbiosintesis

1/7

J Plant Res (2003) 116:233239 The Botanical Society of Japan and Springer-Verlag Tokyo 2003Digital Object Identifier (DOI) 10.1007/s10265-003-0095-5

Springer-VerlagTokyo102650918-94401618-086030669031Journal of Plant ResearchJ Plant Res009510.1007/s10265-003-0095-5

Biosynthesis of cytokinins

JPR SYMPOSIUM

Received: March 6, 2003 / Accepted: March 13, 2003 / Published online: April 29, 2003

Tatsuo Kakimoto

T. Kakimoto

1

(*

)Department of Biology, Graduate School of Science, OsakaUniversity, Toyonaka, Osaka 560-0043, JapanTel. +81-6-68505420; Fax +81-6-68505421e-mail: [email protected]

Present address:

1

Precursory Research for Embryonic Science and Technology

(PRESTO), Science and Technology Corporation, 4-1-8 Honcho,Kawaguchi, Saitama 332-0012, Japan

Abstract

Cytokinins are adenine derivatives with an iso-prenoid side chain and play an essential role in plant devel-opment. Plant isopentenyltransferases that catalyze thefirst and rate-limiting steps of cytokinin biosynthesis haverecently been identified. Unlike bacterial enzymes, whichcatalyze the transfer of the isopentenyl moiety from dime-thylallyldiphosphate (DMAPP) to the

N

6

position of ade-nosine 5

-monophosphate (AMP), plant enzymes catalyzethe transfer of the isopentenyl moiety from DMAPP pref-erentially to ATP and to ADP. The isopentenylated sidechain is hydroxylated to form zeatin-type cytokinins. Analternative pathway, in which a hydroxylated side chain isdirectly added to the

N

6

position of the adenine moiety, hasalso been suggested.

Key words

Arabidopsis

AtIPTs Cytokinins DMAPP :ATP/ADP isopentenyltransferase Plant hormone

Introduction

This review focuses on the biosynthesis of cytokinins. Nat-urally occurring cytokinins are adenine derivatives with aside chain at the

N

6

position. Depending on the structureof the side chain, cytokinins are classified as isoprenoid oraromatic cytokinins, although aromatic cytokinins are rare.An isoprenoid cytokinin is either an isopentenyladenine(iP)-type cytokinin, which carries an isopentenyl

N

6

sidechain, or a zeatin-type cytokinin, which carries hydroxy-

lated isopentenyl

N

6

side chain. The side chain of a zeatin-type cytokinin occurs in either the

cis

or

trans

configuration,depending on which of the two methyl groups of the sidechain is hydroxylated. The activity of

trans

-zeatin is muchhigher than that of

cis

-zeatin. Cytokinins occur in base,riboside, or ribotide forms, with active cytokinins thoughtto be in the base form. Yamada et al. (2001) proved this byshowing that iP and

trans

-zeatin, but not their ribosides,could bind to the cytokinin receptor CRE1/WOL/AHK4.Many modifications of cytokinins are known; Mok and Mok(2001) have reviewed these in detail. Cytokinin oxidases/dehydrogenases destroy cytokinins by cleaving the sidechain (Houba-Herin et al. 1999; Morris et al. 1999).

Levels of cytokinins are spatially and temporally regu-lated. Cytokinins are abundant in the root tip, the shootapical meristem, and immature seeds (Letham 1994). It isgenerally assumed that the root tip is the major site ofcytokinin synthesis, but the cambium, the shoot apex, andimmature seeds are also thought to synthesize cytokinins(Letham 1994; Emery et al. 2000). Changes in cytokininlevels in association with plant development have beenreported (Morris et al. 1993; Benkova et al. 1999; Dewitteet al. 1999; Emery et al. 2000; Yang et al. 2001; Jacqmard etal. 2002). Changes in cytokinin levels are also associatedwith the cell cycle, the levels being highest in the late Sphase and during the M phase (Redig et al. 1996b). Envi-

ronmental factors also affect cytokinin levels, which aregenerally positively correlated with levels of mineral nutri-ents, especially nitrogenous nutrients (Goring and Mar-danov 1976; Salama and Wareing 1979; Samuelson andLarsson 1993; Takei et al. 2001b, 2002; Sakakibara and Takei2002), and decreased by water stress (Yang et al. 2001). Thelevels of active cytokinins in plants are expected to be reg-ulated by the rates of biosynthesis, inter-conversion, trans-port, and degradation.

Until recently, the rate-limiting step of cytokinin biosyn-thesis was assumed to be the addition of the isopentenylside chain to AMP. Recently, however, cytokinin biosyn-thetic isopentenyltransferases have been identified (Kakim-

oto 2001; Takei et al. 2001a), and analysis of these enzymeshas revealed that cytokinins are most likely synthesized by


2/7

234

isopentenylation of ATP and ADP (Kakimoto 2001). Analternative pathway, perhaps involving the addition of ahydroxylated side chain to the adenine moiety, has alsobeen proposed (Astot et al. 2000). Another possible sourceof cytokinins is tRNA, since particular tRNA species areisopentenylated at an adenosine residue. Because this field

is advancing rapidly, it is timely to summarize our currentknowledge about cytokinin biosynthesis.

DMAPP : AMP isopentenyltransferase of

Dictyostelium discoideum

Taya et al. (1978) demonstrated that a partially purifiedenzyme sample from

Dictyostelium discoideum

catalyzedthe transfer of the isopentenyl moiety from dimethylallyldiphosphate (DMAPP) to AMP; this was referred to asDMAPP : AMP isopentenyltransferase activity. ATP, ADP,

and cAMP did not function as isopentenyl acceptors. Thiswas the first demonstration of cytokinin biosynthesis invitro. In

D. discoideum

cells, the reaction product, isopen-tenyladenosine-5

-monophosphate (iPMP), is converted toiP, and iP is then modified at its

N

3

position to form discad-enine, a spore germination inhibitor. Interestingly, discade-nine shows cytokinin activity in assays of tobacco callusgrowth (Nomura and Tanaka 1977), but researchers havenot examined in detail whether isopentenyladenine itself

also has a direct biological role in the development of

D.

discoideum

. It would be interesting to examine which pro-cesses in the development of

D. discoideum

are affectedwhen the gene for the putative DMAPP : AMP isopente-nyltransferase (see Fig. 1) is disrupted. Extracts of

D. dis-

coideum

also possess cytokinin oxidase activity (Armstrong

and Firtel 1989). The identification of AMP-isopentenyl-transferase activity in

D. discoideum

encouraged investiga-tors to examine similar activity in cytokinin-producingbacteria and in plants.

DMAPP : AMP isopentenyltransferases of

Agrobacterium tumefaciens

The first cytokinin biosynthetic enzyme to be identifiedcame from the gall-forming bacterium

Agrobacterium tume-

faciens

(Akiyoshi et al. 1984; Barry et al. 1984). Tissue-

cultured

A. tumefaciens

-incited galls are autotrophic forauxin and cytokinin, even after the gall tissue is cured of

A. tumefaciens

(Brown 1958). When

A. tumefaciens

infectsa plant, the T-DNA region of the Ti-plasmid is introducedinto the plant chromosome. T-DNA carries genes that areexpressed in a plant cell and are responsible for deregulatedproduction of auxin and cytokinin and for tumor formation.The

tms

locus is composed of two genes responsible for theproduction of auxin, and the

tmr

locus consists of a gene

Fig. 1.

A phylogenetic tree forconserved regions (regionscorresponding to 6112 aminoacid residues ofisopentenyltransferases.CLUSTALW program(http://www.ddbj.nig.ac.jp/)was used


3/7

235

responsible for the production of cytokinin. The

tmr

(

ipt

)gene was cloned and expressed in

Escherichia coli

, andextracts of

E. coli

were shown to catalyze production ofiPMP from DMAPP and AMP (Akiyoshi et al. 1984; Barryet al. 1984). The purified

tmr

gene product isopentenylatedAMP, but not ATP or ADP, and is DMAPP : AMP isopen-

tenyltransferase (Morris et al. 1993).Nopaline-producing strains of

A. tumefaciens

possessanother gene for DMAPP : AMP isopentenyltransferase,

tzs

,which is present in the Ti-plasmid outside the T-DNA. The

tzs

gene is responsible for the high level of cytokinin produc-tion by these

A. tumefaciens

strains (Morris et al. 1993).Genes that resemble

tmr

and

tzs

are also present in othergall-forming bacteria, including

Pseudomonas syringae

pv.Savastanoi (Powell and Morris 1986),

Pseudomonas

solanacearum

(

Ralstonia solanacearum

) (Akiyoshi et al.1989), and

Pantoea agglomerans

(

Erwinia herbicola

) (Lich-ter et al. 1995), and in the phytopathogenic bacterium

Rhodococcus fascians

, which causes leaf deformation,witches broom, green patches on laminae, or leafy galls(Crespi et al. 1992; Goethals et al. 2001). These

tmr/tzs

-related genes are also responsible for cytokinin productionin these bacteria (Akiyoshi et al. 1987). Interestingly, notonly phytopathogenic bacteria, but also the nitrogen-fixingsymbiotic cyanobacterium

Nostoc

possesses a gene relatedto

tmr/tzs

(Fig. 1).

Cytokinin biosynthetic isopentenyltransferasesof plants

DMAPP : AMP isopentenyltransferase activity wasdetected in partially purified enzyme samples fromcytokinin-autotrophic cultured cells of tobacco (Chen andMelitz 1979) and from kernels of

Zea mays

(Blackwell andHorgan 1994). In those studies, DMAPP and AMP, eitherof which was radiolabeled, were incubated with the enzymesamples. After the reaction, the mixtures were treated witha phosphatase and then incorporation of radioactivity intothe isopentenyladenosine (iPAdo) fraction was determinedas the enzyme activity. However, since then, no significantprogress in identifying the corresponding enzyme has beenmade. Recently, two groups have identified cytokinin bio-

synthetic isopentenyltransferases by exploiting the

Arabi-dopsis thaliana

genome database (Kakimoto 2001; Takei etal. 2001a). It had been assumed that the first step in cytoki-nin biosynthesis was isopentenylation of AMP; however, astandard BLAST search of the database using bacterialcytokinin biosynthetic AMP isopentenyltransferases as aquery returned plant sequences with only weak resem-blances. Nonetheless, it was still reasonable to think thatsome kind of isopentenyltransferase was involved in a keystep of cytokinin biosynthesis. Therefore, I extracted aminoacid residues that were common to both bacterial DMAPP :AMP isopentenyltransferases and DMAPP : tRNA isopen-tenyltransferases (see below for DMAPP : tRNA isopente-

nyltransferases) (Kakimoto 2001). The common patternwas as follows:

G

x

T

xx

GK[ST]

xxxxx

[VLI]

xxxxxxx

[VLI][VLI]

xx

D

xx

Q

x

{57,60}[VLI][VLI]

x

GG[ST]

where

x

denotes any amino acid residue, [ ] denotes any oneof the amino acid residues within [ ], and

x

{

m

,

n

} denotesamino acid residues of

m

to

n

in number.

Because the common amino acid residues were pre-served in these two types of isopentenyltransferases withdifferent substrate specificities, I thought it possible that thepattern would be present even in isopentenyltransferasesfor unknown substrates. To test this idea, I used thePatmatch program (http://www.arabidopsis.org/cgi-bin/patmatch/nph-patmatch.pl). Unlike the standard BLASTsearch program, the Patmatch program can calculatehomology values for a pattern of specified amino acid res-idues and ignore intervening, unspecified (nonconserved)amino acid residues. In 1999, the database contained twogenes (

AtIPT2

and

AtIPT4

) for products that matched theconserved pattern. When

AtIPT4

was overexpressed in

Arabidopsis

calli, transformed calli underwent typical cyto-kinin responses rapid cell division and shoot formation in the absence of cytokinins.

AtIPT2

, which was later shownto be a gene for DMAPP : tRNA isopentenyltransferase(Golovko et al. 2002), did not have such an effect. Next,biochemical properties were examined. A crude extractof

E. coli

expressing

AtIPT4

exhibited apparent DMAPP :AMP isopentenyltransferase activity: when DMAPP andradiolabeled AMP were incubated with an enzyme sampleand the reactant then treated with a phosphatase, incorpo-ration of radioactivity into iPAdo was detected. However,this activity was an artifact: radiolabeled AMP is readilyconverted to radiolabeled ATP and ADP by the action of

E. coli

-derived factors. Purified AtIPT4 isopentenylatedATP and ADP, but not AMP; therefore, the enzyme shouldbe classified as DMAPP : ATP/ADP isopentenyltrans-ferase. The completed

Arabidopsis

genome database pos-sesses nine genes for putative isopentenyltransferases,which are designated

AtIPT19

. Among them,

AtIPT2

(Golovko et al. 2002), and probably

AtIPT9

, code forDMAPP : tRNA isopentenyltransferases. Takei et al.(2001a) also searched the genome database and identified

AtIPT18

.

E. coli

expressing

AtIPT1

,

3

,

4

,

5

,

6

,

7

, or

8

secreted iP and

trans

-zeatin into the culture medium, andcrude extracts from these

E. coli

exhibited DMAPP : AMP

isopentenyltransferase activity. Examining the biochemicalproperties of purified AtIPT1 showed that although AtIPT1catalyzed DMAPP : AMP isopentenyltransferase activity,the

K

M

values were rather high: 185 m

M for AMP and50 m

M for DMAPP (Takei et al. 2001a). These values aremuch higher than the

K

M

values of the

tmr

product of

Agrobacterium

, which are 85.6 nM for AMP and 8.28 m

Mfor DMAPP (Blackwell and Horgan 1993), and of the

tzs

product of

Agrobacterium

, which are 11.1 m

M for AMP and8.2 m

M for DMAPP (Morris et al. 1993). Indeed, AtIPT1,as well as AtIPT4, has been reported to use ATP and ADPmuch more efficiently than AMP, and the main activity ofAtIPT1 is also that of DMAPP : ATP/ADP isopentenyl-

transferase (Kakimoto 2001). The

K

M

values of AtIPT4were 18 m

M for ATP and 6.5 m

M for DMAPP (Kakimoto


4/7

236

2001). These results, together with the general agreementthat the cellular concentration of ATP is much higher thanthat of AMP the typical ATP/AMP ratio is about 100(Hardie et al. 1998) suggested that the first and rate-limiting steps of cytokinin biosynthesis are mainly the iso-pentenylation of ATP and ADP. A proposed cytokinin

biosynthetic pathway is presented in Fig. 2. The isopenteny-lated products of ATP and ADP are isopentenyl ATP(iPTP) and isopentenyl ADP (iPDP), respectively. Laloueet al. (1974) detected iPTP in iP-fed plants; however, thepresence of the diphosphate and triphosphate forms ofcytokinins has since been ignored in studies that examinedspecies and contents of cytokinins in plants. Recently, zeatindiphosphate and zeatin triphosphate have also beendetected in normal plants (P. Moritz and G. Sandberg,personal communication), supporting the ATP- and ADP-derived pathway.

A phylogenetic tree revealed three branches:DMAPP : AMP isopentenyltransferases, DMAPP : tRNAisopentenyltransferases, and DMAPP : ATP/ADP isopente-nyltransferases (Fig. 1). The DMAPP : ATP/ADP isopente-nyltransferase branch is composed of only plant sequences,suggesting that this family diverged after the plant kingdomappeared. Most gene products in the DMAPP : AMP iso-pentenyltransferase branch belong to phytopathogenic bac-teria, but interestingly there are also gene products of D.discoideum and a cyanobacterium, Nostoc.

Alternative pathway: addition of a hydroxylatedside chain

It has generally been thought that the isopentenyl side chainis further hydroxylated to form zeatin-type cytokinins.

However, recent studies have suggested that in nopaline-producing Agrobacterium strains (Krall et al. 2002), andpossibly also in tms-transformed and wild-type Arabidopsis,zeatin-type cytokinins can also be formed by an alternativepathway. Astot et al. (2000) fed tms-transformed Arabidop-

sis with deuterium oxide (2H2O) and [2H6] iPAdo (each

of the two methyl groups of the side chain carried threedeuterium atoms) simultaneously. The majority of iPMPextracted from the fed plants contained six deuteriumatoms, indicating the conversion from fed [2H6] iPAdo. IfiPMP were subsequently hydroxylated, [2H5]-zeatin ribo-side 5-monophosphate (ZMP) would be formed. However,ZMP with 04 deuterium atoms predominated, indicatingthat an iPMP-independent pathway predominated in thetms-transformants. Therefore, Astot et al. (2000) proposedthe possibility that trans-zeatin is formed by the direct addi-tion of a hydroxylated side chain, with 4-hydroxy-3-methyl-2-(E)-butenyl diphosphate (HMBDP) as the side-chaindonor. Similarly, wild-type Arabidopsis was also shown tohave an iPMP-independent pathway, although the iPMP-dependent pathway also plays an important role (Astot et

Fig. 2. A model for cytokininbiosynthesis in plants. Cytokininsthat directly bind to cytokinin

receptors are shaded


5/7

237

al. 2000). HMBDP is an intermediate of the deoxyxylulose(DXP) pathway of DMAPP and D3-isopentenyldiphosphate(IPP) synthesis, and is formed by the action of the IspGgene product in the plastid (Fig. 3) (Hecht et al. 2001). Thispathway is present in bacteria and chloroplasts, and a fullset of genes for the enzymes required for the pathway ispresent in Arabidopsis (Rohdich et al. 2001). DMAPP isalso synthesized by the mevalonate (MEV) pathway, whichdoes not involve HMBDP. Conversion from iPMP to trans-zeatin, which is involved in the iPMP-dependent pathway,was completely inhibited by metyrapone, an inhibitor ofcytochrome P450. Enzyme activity that hydroxylates theside chains of iP and iPAdo, a process which requiresNADPH, was detected in a microsomal fraction of cauli-flower (Chen and Leisner 1984).

As the names tzs (trans-zeatin secretion) ofAgrobacte-rium tumefaciens and ptz (Pseudomonas trans-zeatin pro-ducing) of P. syringae indicate, bacteria expressing thesegenes secrete trans-zeatin (Beaty et al. 1986; Powell andMorris 1986; Akiyoshi et al. 1987). This can be explainedeither by the possible presence of an isopentenyladeninehydroxylase, or by the presence of an iPMP-independentpathway. Recently, purified tzs protein was shown to cata-lyze the transfer of a hydroxylated side chain from HMBDPto AMP, producing ZMP (Krall et al. 2002), but it is notknown whether plant isopentenyltransferases also use

HMBDP. Overexpression of the tmrgene in tobacco (Rediget al. 1996a; Faiss et al. 1997; McKenzie et al. 1998) and inArabidopsis (Astot et al. 2000) resulted in large increasesin zeatin-type cytokinins, but only subtle or modestincreases in iP-type cytokinins. By contrast, iP-type cytoki-nins were predominantly increased when an endogenousplant cytokinin biosynthetic isopentenyltransferase geneSHO ofPetunia hybrida (Zubko et al. 2002) or AtIPT8 of

Arabidopsis (Sun et al. 2003) was overexpressed. Theseresults suggest that plant cytokinin biosynthetic isopente-nyltransferases, at least in the case of petunia SHO and

Arabidopsis AtIPT8, prefer DMAPP as the side-chaindonor. Obviously, it should be tested whether plant cytoki-

nin biosynthetic isopentenyltransferases are able to useHMBDP as a side-chain donor.

tRNA as a possible source of cytokinins

In 1966, iPAdo and 6-(3-methylbut-2-enylamino)purine,both of which were known to have cytokinin activity, werefound as constituents of two serine tRNAs. Cytokinin moi-eties occur as a modified adenosine residue immediately 3to the anticodon of tRNAs that recognize the codon UNN(Skoog and Armstrong 1970). Cytokinin moieties in tRNAhave been found in virtually all organisms tested, and planttRNAs also contain iPAdo, cis-zeatin riboside (cis-ZR),trans-ZR, 2-methylthio-iPA, and 2-methylthio-ZR (Horgan1984). It is generally thought that these tRNAs are firstisopentenylated and that the isopentenyl side chain may bethen further modified (Cherayil and Lipsett 1977). It ispossible that degradation products of cytokinin-containingtRNAs are sources of cytokinins in plants. However,tRNAs are estimated to account for at most 40% of thecytokinin biosynthesis when calculated from the tRNAturnover rate, and it is generally assumed that tRNAs playonly a minor role, if any, for cytokinin precursors (Barneset al. 1980).

Perspectives

AtIPTs, or at least AtIPT1 and AtIPT4, preferentially iso-pentenylate ATP and ADP. For the complete understand-ing of the biosynthetic route of cytokinins, substratespecificities of all isopentenyltransferases ofArabidopsisshould be examined. Also, it is an open question whetherdirect addition of a hydroxylated side chain to form trans-zeatin occurs in plants. Another important question iswhether AtIPT1, 3, 4, 5, 6, 7, and 8 are indeed responsiblefor the production of major part of cytokinins in plants,because we can not exclude a t-RNA derived pathway oran unidentified pathway for cytokinin biosynthesis. Theonly means to answer this question will be analyses of

knockout plants for AtIPTs. Finally, in order to understandhow cytokinins regulate plant development, we have to

Fig. 3. An outline of themevalonate (MEV) pathway andthe deoxyxylulose (DXP)pathway for isoprenoidbiosynthesis. Possible side-chaindonors for cytokinins are shaded.IPP, D3 isopentenyldiphosphate.

MEP, 2-C-methyl-D-erythritol4-phosphate


6/7


7/7

ckbiosintesis

Documents